Re: [PATCH v4 7/7] x86/vdso: Move out the CPU number store
On Mon, 25 Jun 2018, H. Peter Anvin wrote: > On 06/25/18 08:40, Thomas Gleixner wrote: > > On Fri, 22 Jun 2018, Andy Lutomirski wrote: > > > >> On Fri, Jun 22, 2018 at 11:02 AM Bae, Chang Seok > >> wrote: > >>> > Still setup_cpu_number() sucks. > >>> > >>> How about setup_cpu_and_node_number()? > >> > >> setup_fast_cpu_node_nr()? setup_cpunr_regs()? I don't have any > >> brilliant ideas. Thomas? > > > > It's really hard to find a good one. setup_cpu_node_cache() might work, but > > it's not brilliant either. > > > > Since this applies to the getcpu(2) system call (in addition to the > kernel usage), perhaps setup_getcpu()? Works for me.
Re: [PATCH v4 7/7] x86/vdso: Move out the CPU number store
On 06/25/18 08:40, Thomas Gleixner wrote: > On Fri, 22 Jun 2018, Andy Lutomirski wrote: > >> On Fri, Jun 22, 2018 at 11:02 AM Bae, Chang Seok >> wrote: >>> Still setup_cpu_number() sucks. >>> >>> How about setup_cpu_and_node_number()? >> >> setup_fast_cpu_node_nr()? setup_cpunr_regs()? I don't have any >> brilliant ideas. Thomas? > > It's really hard to find a good one. setup_cpu_node_cache() might work, but > it's not brilliant either. > Since this applies to the getcpu(2) system call (in addition to the kernel usage), perhaps setup_getcpu()? -hpa
Re: [PATCH v4 7/7] x86/vdso: Move out the CPU number store
On Fri, 22 Jun 2018, Andy Lutomirski wrote: > On Fri, Jun 22, 2018 at 11:02 AM Bae, Chang Seok > wrote: > > > > > Still setup_cpu_number() sucks. > > > > How about setup_cpu_and_node_number()? > > setup_fast_cpu_node_nr()? setup_cpunr_regs()? I don't have any > brilliant ideas. Thomas? It's really hard to find a good one. setup_cpu_node_cache() might work, but it's not brilliant either. Thanks, tglx
Re: [PATCH v4 7/7] x86/vdso: Move out the CPU number store
On Fri, Jun 22, 2018 at 11:02 AM Bae, Chang Seok wrote: > > > Still setup_cpu_number() sucks. > > How about setup_cpu_and_node_number()? setup_fast_cpu_node_nr()? setup_cpunr_regs()? I don't have any brilliant ideas. Thomas?
RE: [PATCH v4 7/7] x86/vdso: Move out the CPU number store
> Still setup_cpu_number() sucks. How about setup_cpu_and_node_number()? Thanks, Chang
Re: [PATCH v4 7/7] x86/vdso: Move out the CPU number store
On Fri, 22 Jun 2018, Andy Lutomirski wrote: > > On Jun 22, 2018, at 8:18 AM, Thomas Gleixner wrote: > > > >> On Wed, 20 Jun 2018, Chang S. Bae wrote: > >> > >> +#ifdef CONFIG_X86_64 > >> +static void setup_cpu_number(int cpu) > > > > This function name is really misleading. vdso_setup_cpu_and_node() or > > something like that clearly tells what this is about. > > > Ah. The big secret is that it’s not just vDSO, or at lease hpa wants to > start using it in the entry code. Fair enough. Still setup_cpu_number() sucks. Thanks, tglx
Re: [PATCH v4 7/7] x86/vdso: Move out the CPU number store
> On Jun 22, 2018, at 8:18 AM, Thomas Gleixner wrote: > >> On Wed, 20 Jun 2018, Chang S. Bae wrote: >> >> +#ifdef CONFIG_X86_64 >> +static void setup_cpu_number(int cpu) > > This function name is really misleading. vdso_setup_cpu_and_node() or > something like that clearly tells what this is about. > Ah. The big secret is that it’s not just vDSO, or at lease hpa wants to start using it in the entry code. > Thanks, > >tglx
Re: [PATCH v4 7/7] x86/vdso: Move out the CPU number store
On Wed, 20 Jun 2018, Chang S. Bae wrote: > > +#ifdef CONFIG_X86_64 > +static void setup_cpu_number(int cpu) This function name is really misleading. vdso_setup_cpu_and_node() or something like that clearly tells what this is about. Thanks, tglx
